In late January 2026 a mass migration of families from Pakistan’s Tirah Valley a remote and mountainous enclave in the Khyber materialised, two opposite stories. One of civilians worried moving through the snow trying to reach provisional registration centres, while another comprised of online messages tried to determine how to shape global perceptions of their plight. The first is a humanitarian crisis: Chanceholders in Tirah were registered by authorities to comprise some 10,000 families, or 70,000 people, or nearly half of the total population of the Valley, who fled amid fear of a coming security operation.
The second story is digital versions of events on X, the platform formerly Twitter, amplified grievance discourses, versions of events that significantly contradicted official and local reports, and versions of events that recycled grievance discourses, also, according to several open source investigators and regional monitors consulted for this piece, spreading coded rhetoric that lowers the threshold to radicalisation. At the heart of this cacophony of an ecosystem lies an account known to be regionally referred to as Nukta, an alias, which researcher investigation teams assert is playing a role in a repeatedly associated manner, granting itself content that glorifies violence and casts Pakistan’s security institutions in adverse light.
The issue with this account goes far beyond any one post. Rather, it represents a platform whose structure of moderation has been undermined at the precise moment that an information war has been intensified across, and around, Pakistan, with serious ramifications for civilians population.
The Human Cost: Numbers and Testimony
Relief authorities and local administrators say there are tens of thousands of displaced people, and many of them in neighbouring Bara where they have come with just their bags. Registration efforts-which have been expanded by registering officials in light of the magnitude of the crisis-have been hard pressed to keep up with the magnitude of both humanitarian need and the political discussion about the cause of the movement. Provincial leaders publicly challenged the federal claim that the displacement was “voluntary,” since it put deadlines and mosque announcements and an atmosphere of coercion against many families who were forced to the point of remaining as an option.
The videos and testimonies circulating online-party collections to which the BBC is subscribed-show families, brown, reduced, shivering, the expression of fear and cold, the sudden loss of fire and livelihood, women and children testify to the human element of its history that does not allow the option to make out the exodus as a choice. These raw accounts of human experiences are in competition in the digital world with social media posts that reinterpret the displaced as “foreign infiltrators”, or describe the clearance of the valley in neutral if not celebratory terms. The result is therefore a disputed truth, the people are fleeing; what the world believes to be the driving force of the exodus is becoming more and more determined by feed algorithms.
Nukta as a Reference Point: Messaging, Tone, and Reach
The account, known as Nukta, does not function in an open recruitment channel, but uses strategies that fit the modern picture of influence operations, such as constant compaction of the state as ‘illegitimate’; romanticisation or normalisation of violent resistance, and selective amplification of videos and claims against public empathy with displaced civilians. At the time by research in South Asia online researchers informed the author of Nukta of which they felt used content that is consistent with the message that members of militant actors have been using in the past: delegitimise the ‘state’, justify armed groups and undermine people’s sympathy for victims when used for political purposes. Analysts of the phenomenon say that messaging like this makes it dangerous exactly because it falls somewhere between grievance messaging and political speech not automatically taken down, as well as nudging the audience into accepting violence.
Geolocation and temporal analyses by independent open source investigators, and here complicating and, as always with the necessary caveats concerning the assignment of statistical probabilities to the results, suggest behavioural markers which are consistent with the activity originating outside of Pakistan. These include post-timing patterns and cross network amplification from accounts previously associated with a larger or Central Asian language influencer space. Whether adopting these patterns points to direct state sponsorship, diaspora activism or hidden more covert networks of militants, the immediate effect is the same: a platform for narratives for which the immediate point of inference is to exacerbate an already fraught security environment.
Platform Responsibility: X Eroded Moderation Engine
Understanding how one account can have a disproportionate effect involves the consideration of the context of the platform. Since the acquisition by Elon Musk, X has made a dramatic change in their approach to content moderation spending a lot less time on trust & safety staffing, shifting towards more of a report driven strategy and repeatedly restructuring their regional capabilities. Multiple outlets have recorded sustained lost moderation capacity who’s staff decreased X is now left with a fraction of the principle or moderation bench that was enjoyed by larger teams.
These shifts in structure have tangible consequences. Messaging of an extremist and grievance-based nature spread fatser than it can be contained. Signals in non English language, coded rhetoric and synchronised micro-amplification are harder to detect without regional teams and also pro-active threat intelligence. Organisations which monitor the violent extremist use of tech platforms said last year the risk environment was escalating throughout south Asia and central Asia. Groups are “testing” narratives openly on mainstream platforms before bringing audiences into more closed channels.
Disinformation as a Force Multiplier
Disinformation is not an abstract problem, but rather a force multiplier in conflict zones. In the case of Tirah, where reporting that is often foggy while operations are also secretive and political blame is already colliding, the curations of selective facts and the correlating of inflammatory framing has a powerful political effect: which can be delegitimizing humanitarian appeals, increase hardening of public attitudes, and even justify further securitized responses under the guise of “clearing out foreign elements.”
Several independent fact-checking platforms have reviewed posts from the Nukta account and publicly debunked them as disinformation.
What X Could Do and What It Hasn’t
Platform policy, in most cases, already forbids praise and support for extremist violence and coordinated disinformation campaigns. The weak link is enforcement: if a narrative is coded as political grievance or framed in ambiguous language, automated systems and lean teams are less likely to act. Public reporting suggests X has not provided clear transparency on how regionally specific threats and language-based radicalization are prioritized and, in the case of Nukta and similar handles, civil-society reporters say they have encountered little public explanation when accounts remain active despite repeated reporting.
Accountability here requires three things, better regional language moderation and pro-active threat intelligence, as well as transparent disclosure when anonymized accounts used to spread extremist narratives have been investigated or left active. That transparency must include public metrics showing how platforms are tackling violent radicalization in non-English contexts. Until platforms comprehend those as fundamental functions with respect to safety, not ancillary costs, the information ecology that informed Tirah’s crisis will continue to pose a threat multiplier. The voices of those who fled will continue to be battled over, part of an online conflict that they never sought to enter.
